Cell-specific Regulation of PTX3 by Glucocorticoid Hormones in Hematopoietic and Nonhematopoietic Cells*

PTX3 (prototypic long pentraxin 3) is a fluid phase pattern recognition receptor, which plays nonredundant roles in the resistance against diverse pathogens, in the assembly of a hyaluronic acid-rich extracellular matrix, and in female fertility. Inflammatory signals induce production of PTX3 in diverse cell types, including myeloid dendritic cells (DC), fibroblasts, and endothelial cells (EC). The present study was designed to explore the effect of glucocorticoid hormones (GC) on PTX3 production in different cellular contexts. In myeloid DC, GC inhibited the PTX3 production. In contrast, in fibroblasts and EC, GC alone induced and, under inflammatory conditions, enhanced and extended PTX3 production. In vivo administration of GC augmented the blood levels of PTX3 in mice and humans. Moreover, patients with Cushing syndrome had increased levels of circulating PTX3, whereas PTX3 levels were decreased in subjects affected by iatrogenic hypocortisolism. In nonhematopoietic cells, GC receptor (GR) functioned as a ligand-dependent transcription factor (dimerization-dependent) to induce PTX3 gene expression. In contrast, in hematopoietic cells, GR repressed PTX3 gene transcription by interfering (dimerization-independent) with the action of other signaling pathways, probably NFκB and AP-1. Thus, divergent effects of GC were found to be due to different GR mechanisms. The results presented here indicate that GC have divergent effects on PTX3 production in hematopoietic (DC and macrophages) and nonhematopoietic (fibroblasts and EC) cells. The divergent effects of GC on PTX3 production probably reflect the different functions of this multifunctional molecule in innate immunity and in the construction of the extracellular matrix.


PTX3 (prototypic long pentraxin 3) is a fluid phase pattern recognition receptor, which plays nonredundant roles in the resistance against diverse pathogens, in the assembly of a hyaluronic acid-rich extracellular matrix, and in female fertility. Inflammatory signals induce production of PTX3 in diverse cell types, including myeloid dendritic cells (DC), fibroblasts, and endothelial cells (EC).
The present study was designed to explore the effect of glucocorticoid hormones (GC) on PTX3 production in different cellular contexts. In myeloid DC, GC inhibited the PTX3 production. In contrast, in fibroblasts and EC, GC alone induced and, under inflammatory conditions, enhanced and extended PTX3 production. In vivo administration of GC augmented the blood levels of PTX3 in mice and humans. Moreover, patients with Cushing syndrome had increased levels of circulating PTX3, whereas PTX3 levels were decreased in subjects affected by iatrogenic hypocortisolism. In nonhematopoietic cells, GC receptor (GR) functioned as a ligand-dependent transcription factor (dimerization-dependent) to induce PTX3 gene expression. In contrast, in hematopoietic cells, GR repressed PTX3 gene transcription by interfering (dimerization-independent) with the action of other signaling pathways, probably NFB and AP-1. Thus, divergent effects of GC were found to be due to different GR mechanisms. The results presented here indicate that GC have divergent effects on PTX3 production in hematopoietic (DC and macrophages) and nonhematopoietic (fibroblasts and EC) cells. The divergent effects of GC on PTX3 production probably reflect the different functions of this multifunctional molecule in innate immunity and in the construction of the extracellular matrix.
Pentraxins are a superfamily of conserved proteins usually characterized by a cyclic multimeric structure (1)(2)(3)(4). The classic short pentraxins C-reactive protein (CRP) 4 and serum amyloid P component (SAP) are acute phase proteins in humans and mice, respectively. They are produced in the liver in response to inflammatory signals, most prominently IL-6. CRP and SAP bind different ligands, in a calcium-dependent manner, and they are involved in the innate resistance to microbes and the scavenging of cellular debris and extracellular matrix components (5).
Recent studies in gene-modified mice have shown that PTX3 plays complex nonredundant functions in vivo, ranging from innate immunity against diverse microorganisms to the assembly of a hyaluronic acid-rich extracellular matrix and female fertility (3, 6, 14, 16, 17, 20 -22). PTX3 acts as a fluid phase pattern recognition receptor, which plays a nonredundant role in resistance against selected pathogens and has properties similar to antibodies. Its production is induced by pathogen recognition, and it recognizes microbial moieties, activates the classical pathway of complement, and facilitates recognition by macrophages and DC. Thus, PTX3 behaves as a bona fide anteantibody (3,6,14,17,18).
PTX3 is also a constituent of the extracellular matrix, which plays a nonredundant role in the assembly of the hyaluronanrich viscoelastic matrix of the cumulus oophorus and hence in female fertility (3,6,14,16,21,23). PTX3 binds TSG6, a hyaluronic acid-binding protein, and may act as a focal point of the assembly of hyaluronic acid-rich matrices (3,6,16).
In addition, PTX3 has a tuning function under inflammatory conditions at least in part limiting tissue damage by regulating apoptotic cell clearance and acting as a third party agent between microbial stimuli and dying cells (3, 19, 22, 24 -27).
Glucocorticoid hormones (GC) have a complex regulatory activity on components of innate and adaptive immunity (28 -30). Since PTX3 is a multifunctional fluid phase pattern recognition receptor that links innate immunity, inflammation, and matrix deposition and shows a unique pattern of regulation (3,12), we explored the GC effect on the PTX3 production. The present study reports that exogenous as well as endogenous GC regulate PTX3 production in a cell type-specific manner and through different actions of GC receptor (GR). The divergent effect of GC on PTX3 production in hematopoietic (DC and macrophages) and nonhematopoietic (fibroblasts and EC) cells mirrors the diverse functions of this molecule in innate immunity and in the assembly of the extracellular matrix.
DC and Macrophages-Monocyte-derived DC were generated as previously described (8). Briefly, blood monocytes were obtained from fresh buffy coats of healthy donors (courtesy of the Centro Trasfusionale, Ospedale di Desio, Milan, Italy) by Ficoll (Sigma) and Percoll (Amersham Biosciences) and after discard of nonadherent cells. Purified monocytes were cultured for 6 days at 1 ϫ 10 6 cells/ml in 6-well tissue culture plates (BD Biosciences) in RPMI 1640 medium supplemented with 2 mM L-glutamine and 10% FCS, 50 ng/ml granulocyte-macrophage colony-stimulating factor, and 20 ng/ml IL-13.
The RAW 264.7 murine macrophage cell line was obtained from ATCC. Bone marrow-derived macrophages were obtained from WT and GR dimerization knock-out (GR dim/dim ) mice, as previously described (31).
Cells were then cultured for 24 h in RPMI 1640 medium supplemented with 2 mM L-glutamine and 2% FCS at 1 ϫ 10 6 cells/ml in 24-well tissue culture plates (BD Biosciences) in the presence of LPS, 1 and 100 ng/ml Dex, ranging from 10 Ϫ5 to 10 Ϫ7 M, and 10 Ϫ5 M RU486. RU486 and Dex were added 1 h prior to the challenge with LPS.
Fibroblasts and EC-Human normal lung fibroblasts, WI38, and the human fibrosarcoma cell line, 8387, were obtained from ATCC. WI38 and 8387 cell lines were cultured in RPMI supplemented with 10% FCS and 2 mM L-glutamine until stimulation.
Mouse embryonic fibroblasts (MEF) from WT and GR dim/dim mice were obtained as previously described (32) and prepared according to standard techniques. Cells were plated at a density of 1 ϫ 10 5 cells and expanded in 6-well plates in RPMI 1640 medium supplemented with 2 mM L-glutamine, 10% FCS, and ␤-mercaptoethanol. The number of MEF at confluence was 5 ϫ 10 5 cells/well. Cells were stimulated at passage 7-9.
Human umbilical vein endothelial cells (HUVEC) were isolated from umbilical cord vein by digestion with 0.5% collagenase, as previously described (33). HUVEC were then plated at a density of 1 ϫ 10 5 cells/well in gelatin-coated 24-well plates (BD Biosciences) and grown to confluence in medium 199/Dulbecco's modified Eagle's medium (1:1) supplemented with 20% FCS, 2 mM L-glutamine, 100 g/ml streptomycin, 100 units/ml penicillin, 5 g/ml endothelial cell growth factor, and 10 g/ml heparin. Cultured medium was refreshed every other day. The number of HUVEC at confluence was 5 ϫ 10 5 cells/well. Confluent cells were then cultured for the reported periods in 24-well tissue culture plates (BD Biosciences) in medium supplemented with 2 mM L-glutamine, 2% FCS in the presence of 20 ng/ml TNF␣ or murine TNF␣, 20 ng/ml IL-1␤, Dex ranging from 10 Ϫ5 to 10 Ϫ8 M, and 10 Ϫ5 M RU486. RU486 and Dex were added 1 h before to challenge with proinflammatory stimuli.
Plasmid Construction, Transfection, and Luciferase Assay-A 1.37-kilobase EcoRI-PvuII genomic fragment from lP2 phage (10), which spans nucleotides 1-1374 of the human PTX3 promoter (accession number X97748), was used as DNA template for PCR with the following oligonucleotides. Forward primer (5Ј-AAAAAGGTACCCCGGATCTCCCTTCTAAC-3Ј) was designed in antisense orientation on the PTX3 promoter from nucleotide 7 to nucleotide 25 with an additional 11 nucleotides at the 5Ј-end; six of them (underlined) give rise to a KpnI restriction site; reverse primer (5Ј-TTTCCACAGCTG-GCGGGAGGAGACTCTCAA-3Ј) was designed in sense orientation on the PTX3 promoter from nucleotide 1351 to 1380; it includes a PvuII site (underlined). The PCR products were subcloned in pGEMT easy vector; next, the human PTX3 promoter fragment was excised by KpnI-PvuII digestion and subcloned in SmaI-KpnI-digested pGV-B2 vector. The luciferase reporter clones containing the human PTX3 promoter were sequenced (Primm) and subsequently used in functional assays. Cell transfections were performed using Lipo-fectamine 2000 TM according to the manufacturer's instructions (Invitrogen).
A luciferase assay was performed using the luciferase assay kit from Promega (catalog number E1500). Briefly, cells were stimulated at various times, as described, washed in PBS, and resuspended in cell lysis buffer (Promega kit). Cell lysate proteins were quantified using the Coomassie method (Bio-Rad). In order to perform the luciferase assay, 50 l of luciferase substrate, reconstituted following the manufacturer's instructions, were added to 15 g of total proteins. Signals was read on Luminometer LB 9507 (Berthold Inc.) RNA-For Northern blot analysis, total RNA was extracted by the TRIzol method according to the manufacturer's instructions (Invitrogen), blotted, and hybridized as described (8). PTX3 mRNA densitometric analysis was performed by quantification of relative optical density ϫ area, using Image Analysis Software (Image Research, Inc.).
For gene expression analysis by quantitative PCR, according to the manufacturer's instructions, total RNA was reverse transcribed using a cDNA archive kit (Applied Biosystems), and then real time PCR was performed with Power Syber Green PCR Master Mix (Applied Biosystems) and detected by a 7900HT fast real time system (Applied Biosystems). Data were processed using the SDS2.2.2 software (Applied Biosystems). All results were normalized to the expression of the housekeeping gene, ␤-actin, in the PCRs and expressed as -fold change in mRNA expression with respect to the untreated cells. Data represented are from three independent experiments done in triplicate.
Chromatin Immunoprecipitation Assay (ChIP)-ChIP assays were performed in HUVEC, WI38, and DC, as previously described (34). Chromatin was sheared by sonication to give fragments 500 -1000 bp in length. Chromatin immunoprecipitation was performed overnight at 4°C with 4 g of a rabbit polyclonal anti-RNA pol II antibody (N-20; catalog number B0408), a rabbit polyclonal anti-GR (H-300; catalog number B0604), and a control antiserum (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Immunoprecipitated DNA was amplified by real time PCR with PTX3 promoter-specific primers spanning the PTX3 promoter from nucleotide 1231 to nucleotide 1380. The ChIP results were normalized to input and expressed as -fold induction with respect to the control cells. Data are from three independent representative experiments done in triplicate.
Proteins-Human PTX3 was measured using an ELISA previously described (35). IL-6 was measured by ELISA in according to the manufacturer's instructions (R&D Systems, Minneapolis, MN).
Murine PTX3 was measured by ELISA based on monoclonal antibodies obtained by immunization of PTX3 Ϫ/Ϫ mice (14). In brief, 96-well ELISA plates (Nunc, Roskilde, Denmark) were coated with 100 ng of anti-murine PTX3 monoclonal antibody MnmE1. Nonspecific binding sites were saturated by 300 l of 5% dry milk, prior to the addition of purified recombinant murine PTX3 standards (from 0.1 to 10 ng/ml) and unknown samples, diluted in PBS (Cambrex Bio Science) plus 2% bovine serum albumin (Sigma). After incubation, 25 ng of biotin-conjugated anti-murine PTX3 monoclonal antibody MNmE2 were added. Streptavidin-peroxidase was then added, followed by the addition of the TMB liquid substrate system (Sigma). After each step, plates were extensively washed with PBS (Cambrex Bio Science) containing 0.05% Tween 20 (Merck), also used for reagent diluent.
Immunohistochemistry-Skin samples (one normal and one with Cushing syndrome) were collected and fixed in 4% formalin. Paraffin-embedded tissue was cut at 4 m and incubated with 2% H 2 O 2 in methanol (to eliminate endogenous peroxidase activity). Sections were preincubated with PBS, 1% BSA, 1% normal human serum, 0.02% Nonidet P-40 (Sigma) to block nonspecific sites for 10 min at room temperature and than incubated at room temperature for 2 h with an anti-PTX3 affinity-purified rabbit polyclonal antibody (0.25 g/ml). A purified rabbit serum was also used as negative control. After extensive wash, sections were incubated with secondary antibodies; the Super Sensitive IHC detection system (BioGenex, San Ramon, CA) was used according to the manufacturer's instructions. Staining reactions were performed with 3,3-diaminobenzidine (BioGenex) as substrate. Sections were counterstained with hematoxylin and mounted with Eukitt.
Mice-PTX3 levels were measured in male, adult (8 weeks), C57/BL6 mice (Charles River Laboratories, Calco, Italy). Dex was prepared as a suspension in PBS (Cambrex Bio Science, Belgium) at 5 mg/ml and was administered, at 30 mg/kg, via intraperitoneal injection prior to blood collection at 24 h. Control mice received intraperitoneal injections of vehicle. Six animals were examined for each experimental group.
Patients-The study included sixteen patients with hypercortisolism and eight patients with hypocortisolism admitted to the Endocrine Unit, Fondazione Policlinico, IRCCS. All patients with hypercortisolism (five males and 11 females, aged 34 -72) were affected by ACTH-dependent Cushing syndrome due to either a pituitary adenoma or to ectopic ACTH syndrome. All patients had elevated 24 h urinary cortisol excretion and lack of suppression of cortisol levels after a low dose Dex test. Patients with hypocortisolism (four males and five females, aged 40 -67) had a history of Cushing disease due to a pituitary adenoma surgically removed by the transsphenoidal route. Three months after surgery, all patients had reduced 24 h urinary cortisol excretion and lack of induction of cortisol levels after a low dose (1 g) ACTH stimulation test. The patients were evaluated after adequate cortisone acetate treatment withdrawal. Serum PTX3 and cortisol levels were measured in seven normal subjects, both under basal conditions and after an overnight treatment with 1 mg of Dex. PTX3 was measured by ELISA as described. Serum cortisol levels were measured by the enzyme immunoassay method (Nichols Institute, San Juan Capistrano, CA). The control cohort was from the same unit. Informed consent was obtained in all cases after project approval by the local ethics committee.

Regulation of PTX3 Production by GC in
Monocyte-derived DC-In a first series of experiments, we evaluated the effect of GC on monocyte-derived DC and their capability to regulate the PTX3 production. Fig. 1A summarizes results of a series of 12 experiments where monocyte-derived DC were treated with PTX3 Regulation by Glucocorticoids OCTOBER 31, 2008 • VOLUME 283 • NUMBER 44

JOURNAL OF BIOLOGICAL CHEMISTRY 29985
LPS (1 and 100 ng/ml), Dex (10 Ϫ5 to 10 Ϫ7 M), and combinations thereof. As expected, LPS induced PTX3 production in DC (8) and resulted in release of 14.8 Ϯ 6.7 and 42.7 Ϯ 12.0 ng/ml (mean Ϯ S.D.) when DC were challenged for 24 h with either 1 or 100 ng/ml LPS, respectively. GC alone did not induce PTX3 in DC. In all experiments performed, Dex caused a strong inhibition of the LPS-induced PTX3 production. As shown in Fig.  1A, when Dex (10 Ϫ5 M) was combined with either 1 or 100 ng/ml LPS, the inhibition of PTX3 production was 63.0 Ϯ 14.5 or 59.5 Ϯ 12.0%, respectively. The inhibitory effect of GC on PTX3 production by monocyte-derived DC was dose-related (range of 10 Ϫ5 to 10 Ϫ7 M) (Fig. 1A).
When RU486 (mifepristone), a GR antagonist, was combined to LPS and Dex, the regulatory effect of GC on PTX3 production by myeloid DC was blocked. As shown in Fig. 1B, in a series of four experiments, PTX3 production induced by LPS (61.1 Ϯ 4.2 ng/ml) was inhibited by the addition of Dex (37.0 Ϯ 3.4 ng/ml) but not when RU486 was added to the culture (57.1 Ϯ 12.1 ng/ml). RU486 did not induce PTX3 and did not alter the LPS-induced PTX3 production by DC (65.1 Ϯ 13.2 ng/ml) (Fig.  1B). Thus, the inhibitory effect of Dex on LPS-induced PTX3 production by myeloid DC is GR-dependent.
PTX3 transcript levels were assessed by Northern blot analysis, and, as shown in Fig. 1C, Dex was found to cause a marked inhibition of PTX3 mRNA expression induced by LPS in DC after 8 h of culture. Densitometric analysis revealed that a combined treatment with LPS and Dex used at 10 Ϫ5 or 10 Ϫ6 M caused an inhibition of transcript levels corresponding to 64 or 59% of the response to LPS, respectively. GC also inhibited the PTX3 production by LPS-treated monocytes and macrophages (data not shown).
Regulation of PTX3 Production by GC in Nonhematopoietic Cells-Since PTX3 is produced by several cell types, including fibroblasts and EC (3), we evaluated also the ability of GC to regulate the PTX3 production in a human fibrosarcoma (8387), normal human fibroblasts (WI38), and HUVEC.
It was important to ascertain whether the production of PTX3 by GC in cells of fibroblastic origin was receptor-dependent. In a series of three experiments (Fig. 2B), PTX3 production by GC-stimulated 8387 (42.0 Ϯ 7.1 ng/ml) was significantly inhibited when cells were cultured in the presence of RU486 (26.1 Ϯ 1.6 ng/ml, p Ͻ 0.05). As expected, TNF␣-induced PTX3 production (93.1 Ϯ 5.4 ng/ml) by 8387 was markedly augmented in combination with Dex (132.8 Ϯ 15.0 ng/ml). The augmentation by Dex was inhibited by the addition of RU486 (86.0 Ϯ 14.6 ng/ml). RU486 did not inhibit the TNF␣-induced PTX3 production (88.4 Ϯ 7.6 ng/ml) (Fig. 2B). Thus, RU486 blocked the effects of GC on PTX3 production both in fibroblasts and in DC.  Similar results were obtained when the WI38 normal fibroblast cell line was used (four experiments performed; a representative shown in Fig. 3, A and B).
EC are a major source of PTX3 (3). Dex induced PTX3 production in HUVEC (19.7 Ϯ 1.7 ng/ml in the experiment shown in Fig. 4A, representative of four experiments performed) although less efficiently than IL-1␤ (28.7 Ϯ 6.2 ng/ml). Dex superinduced PTX3 in combination with IL-1␤, corresponding to 164.7 Ϯ 11.2% of the IL-1␤ response (63.3 Ϯ 5.2 ng/ml) (Fig.  4A). RU486 completely blocked the production of PTX3 by Dex but did not alter the IL-1␤-induced PTX3 production (Fig. 4B). The effect of GC on PTX3 production by 8387 fibroblasts was dose-related (10 Ϫ5 M-10 Ϫ8 M) (Fig. 2C). Vitamin D 3 and prostaglandin E 2 , which inhibit the LPS-induced PTX3 production in myeloid DC (12), did not induce and did not alter PTX3 production in HUVEC (Fig. 5) and in fibroblasts (data not shown). Stimulation with different hormones (e.g. T3 and T4), which, as GC, engage receptors belonging to the superfamily of nuclear receptors, did not affect PTX3 production by DC and fibroblasts (data not shown).
The time course of PTX3 production in nonhematopoietic cells was different in response to TNF␣ and GC ( Fig. 6; three experiments performed). PTX3 induction in response to TNF␣ was already evident after 8 h, whereas induction by GC required 16 h. GC-induced PTX3 production was sustained up to 48 h. Sustained production over time was also observed when TNF␣ and GC were combined (Fig. 6).   GC were found to induce PTX3 mRNA and to enhance the induction by TNF␣ and IL-1␤ in fibroblasts and EC (range of 5-10 h). The addition of RU486 inhibited the GR-mediated augmentation of PTX3 expression (Fig. 7).
Mechanism of PTX3 Regulation by GC-It was important to explore the molecular mechanism of the divergent PTX3 regulation by GC. PTX3 promoter was found to contain binding sites for the transcription factors NFB and AP-1, which confer responsiveness to IL-1␤, LPS, and TNF␣ (36,37). We investigated whether GC directly modulates PTX3 gene expression. The analysis of the human PTX3 gene promoter (GenBank TM accession number X97748) by Signal Scan software (38) allowed us to identify putative GR-responsive regions. Thus, to evaluate GC regulation of PTX3 gene expression in both fibroblasts and macrophages, a reporter plasmid (pGV-B2) carrying the human PTX3 promoter linked to luciferase gene was generated and used to transiently transfect RAW 264.7 and 8387 cell lines. Therefore, we evaluated the transcriptional regulation of the PTX3 promoter by GC in fibroblasts and macrophages.
In the RAW 264.7 cell line PTX3 promoter activity significantly increased (3.2-fold, p ϭ 0.009) following a 16 h challenge with LPS (10 ng/ml) compared with untreated cells. Treatment with Dex (10 Ϫ6 M) counteracted (Ϫ1.79-fold, p ϭ 0.039) the transcriptional activation of the PTX3 promoter induced by LPS (a representative experiment of three performed is shown in Fig. 8A). In RAW 264.7 cells, Dex did not induce PTX3 promoter activity. The GC-mediated PTX3 gene repression was also observed when cells were also treated for 4 and 8 h (not shown). Fig. 8A shows one of three experiments performed, with the other two shown in the supplemental materials (Fig.  S1). In this cellular context, Dex indeed inhibited NFB and AP-1 activation, as shown by Western blot and EMSA analysis (Fig. S2).
The 8387 cell line was treated with TNF␣ (20 ng/ml), Dex (10 Ϫ6 M), and RU486 (10 Ϫ5 M) for different times after 16 h of transient transfection with PTX3 promoter-pGV-B2 reporter plasmid. Fig. 8B shows one representative experiment at 16 h of three performed (see Fig. S3 for the two experiments not shown). The PTX3 promoter responded to both TNF␣ (1.73fold, p ϭ 0.003) and Dex (1.72-fold, p ϭ 0.013) used alone, and its activity was further enhanced following a combined treatment (2.34-fold, p ϭ 0.017). The luciferase activity induced by GC was repressed when cells were incubated in the presence of the antagonist RU486 (Fig. 8B). Similar results were obtained at 4 and 8 h (data not shown).
Following binding to the GR, GC exert diverse regulatory activities through complex mechanisms. The GR either directly engages GC response elements in the regulatory regions of genes, or indirectly it modulates gene expression interfering with the action of other transcription factors (30,39). GR dimerization-deficient (GR dim/dim ) mice provide a tool to differ the mechanism of action of GC, in which GC response elementmediated gene activation, which is entirely dependent on GR dimerization, is removed but GR interactions with NFB and   AP-1, which are independent of dimerization, are still possible (32). Therefore, we ascertained the effect of GC on PTX3 expression in bone marrow-derived macrophages and MEF from WT or GR dim/dim mice.
First, we employed a quantitative PCR to explore the changes in regulation of LPS-induced expression of PTX3 mRNA in bone marrow-derived macrophages from WT and GR dim/dim mice. RNA from macrophages was collected following treatment with LPS (10 ng/ml), Dex (10 Ϫ6 M), or LPS plus Dex for different times. One representative experiment at 8 h of three performed is shown in Fig. 9A, and the other two are presented in Fig. S4. Dex alone was able to appreciably modify the steady state levels of PTX3 mRNA in macrophages from WT and GR dim/dim mice, whereas LPS caused a 20-fold increase in transcript in treated macrophages derived from WT or GR dim/dim mice. PTX3 expression was strikingly down-regulated by Dex both in WT (p ϭ 0.0098) and GR dim/dim (p ϭ 0.0040) macrophages (Fig. 9A). Similar results were obtained at 2, 4, and 24 h (data not shown).
Then MEF from GR dim/dim mice were assessed for their ability to produce PTX3 in response to GC. A series of three exper-iments was performed (Fig. 9B). As expected, MEF from WT mice produced 8.4 Ϯ 1.34 ng/ml (p ϭ 0.003 versus untreated), 30.1 Ϯ 0.6 ng/ml, and 52.6 Ϯ 3.8 ng/ml (p ϭ 0.002 versus TNF␣) PTX3 in response to Dex, TNF␣, and the combined treatment, respectively. PTX3 induction and superinduction by Dex corresponded to 28.0 and 174.6% of the response to TNF␣ (Fig.  9B). In contrast, MEF from GR dim/dim mice did not produce PTX3 in response to Dex (5.9 Ϯ 1.3 ng/ml) compared with 7.0 Ϯ 0.1 ng/ml for untreated cells. In GR dim/dim mice, no significant superinduction of the response to TNF␣ by Dex was observed (Fig. 9B).
Overall, these results indicate that in hematopoietic cells, GC⅐GR complex binding to NFB and AP-1 transcription factors leads to inhibition of PTX3 gene expression, whereas in nonhematopoietic cells, GC⅐GR complexes directly activate PTX3 transcription by GR binding to the PTX3 promoter. To definitely assess the importance of GC⅐GR complexes in the transcriptional activation of the PTX3 gene, ChIP analysis was conducted in WI38 (n ϭ 3) and HUVEC (n ϭ 2) with RNA pol II and GR antibodies. In agreement with gene expression analysis, results obtained in WI38 (one representative experiment is shown in Fig. 10A, top) and HUVEC (Fig. S5) showed that binding of RNA pol II to the PTX3 promoter was induced upon stimulation (2 h) with Dex (10 Ϫ6 M), TNF␣ (20 ng/ml), and the combination of two stimuli. Furthermore, Dex, but not TNF␣, stimulation was able to induce GR recruitment to the PTX3 promoter (Fig. 10A, bottom), thus demonstrating that GC⅐GR complexes play a role in the transcriptional activation of the PTX3 gene in WI38 and HUVEC. In contrast, ChIP analysis performed in hematopoietic cells (DC) (n ϭ 2; one experiment shown) showed that the LPS-induced binding of RNA pol II to the PTX3 promoter was inhibited by Dex (Fig. 10B).
Effect of GC on PTX3 Levels in Mice and Humans-It was of interest to ascertain the in vivo effect of exogenous GC on PTX3 levels in mice. Fig. 11 shows a typical experiment, of three performed, in which C57/BL6 mice following a 24-h treatment with Dex (30 mg/kg) markedly showed increased circulating levels of PTX3 (53.7 Ϯ 6.8 ng/ml, n ϭ 6, p Ͻ 0.05) compared with control animals (22.0 Ϯ 4.4 ng/ml, n ϭ 6).
Circulating levels of PTX3 were then measured in patients affected by Cushing syndrome (hypercortisolism) and hypocortisolism in order to evaluate a regulatory effect of endogenous GC hormones on PTX3 in humans. As shown in Fig. 12A, the PTX3 serum concentration in 18 normal donors was 2.6 Ϯ 0.3 ng/ml, similar to that observed in other studies (33,38). On the other hand, PTX3 levels were augmented in patients affected by Cushing syndrome (6.0 Ϯ 0.6 ng/ml, n ϭ 16, p ϭ 0.00015 versus control group) and significantly decreased in patients affected by hypocortisolism (1.1 Ϯ 0.1 ng/ml, n ϭ 9, p ϭ 0.00051 versus control group). In seven normal subjects, circulating levels of PTX3 were measured 24 h after treatment with a low dose of Dex (1 mg). PTX3 was found to be augmented after Dex administration from 2.2 Ϯ 0.7 to 5.0 Ϯ 1.6 ng/ml, p ϭ 0.0073 (Fig. 12B).
When histological analysis was conducted in the skin of a patient affected by severe Cushing syndrome, strong interstitial staining for PTX3 was observed in the derma (Fig. 12C).

DISCUSSION
GC have profound effects on innate and adaptive immune responses and are a major tool in the armamentarium of antiinflammatory/immunosuppressive agents (28 -30). In view also of the use of PTX3 as a novel marker of inflammation (3, 40 -42), it was important to assess the effect of GC on this key component of humoral innate immunity. Unexpectedly, we found that GC have divergent effects on PTX3 production in hematopoietic (macrophages and DC) and nonhematopoietic (fibroblasts and EC) cells, with inhibition in the former and induction in the latter.
Cell context-dependent regulation was also observed when the PTX3 promoter was tested in a reporter gene assay with    18). B, PTX3 levels in seven normal subjects treated with Dex (1 mg) before obtaining peripheral blood (overnight treatment). C, histological analysis of PTX3 localization in skin of a patient affected by Cushing syndrome and a healthy donor are shown. PTX3 staining was performed as described under "Experimental Procedures." Brown, PTX3; bar, 50 m. divergent regulation in macrophages (inhibition) and fibroblastic cells (induction). Inhibition and induction are both GR-mediated, as revealed by a specific antagonist (RU486), but ordered by different mechanisms of action.
The GR exerts regulatory activity through complex mechanisms; it functions as a ligand-dependent transcription factor through direct (dimerization-dependent) DNA binding, or it represses gene transcription through a protein-protein (dimerization-independent) interference with the action of other signaling pathways (30,39,43,44).
Here we report that GC-mediated down-regulation of PTX3 in hematopoietic cells is independent of GR dimerization, since LPS-induced PTX3 expression was efficiently down-regulated in macrophages from GR dim/dim mice. On the other hand, PTX3 induction was observed in fibroblasts from WT mice but not GR dim/dim mice.
GC exert many anti-inflammatory actions by interfering with other signaling pathways, and GR interactions with NFB and AP-1 occur in GR dim/dim mice, in which GR-mediated gene activation is impaired (32,45). The human and murine PTX3 promoters have both been characterized, and NFB plays a pivotal role in PTX3 gene regulation (36,37). More recently, PTX3 has been shown to be regulated by the FUS/CHOP oncogene (46,47). Here we report that the human PTX3 promoter contains putative GR-responsive elements, that in nonhematopoietic cells the PTX3 promoter is activated by GC, and that the stimulation of PTX3 gene expression and production is dimerization-dependent. In contrast, GC-mediated suppression of PTX3 expression in cells of hemopoietic origin is dimerization-independent and probably mediated by interference with the NFB pathway. Chromatin immunoprecipitation experiments confirmed the transcriptional activation of the PTX3 gene by Dex in nonhematopoietic (fibroblasts and endothelial cells) but not in myeloid cells. The results offer new unexpected vistas on cell context-dependent PTX3 gene regulation.
The finding that the effect of GC on PTX3 production is cell context-dependent is hardly surprising. Data presented here establish that PTX3 production is subjected to regulatory constraints that differ according to cell type. The phenomenon of GC-mediated opposite cell-specific effects on the apoptotic pathway has been demonstrated by different authors. For instance, although GC mediate proapoptotic events in the majority of hematological cells, GC-induced antiapoptotic signaling has been identified in cells of epithelial origin, and the underlying mechanisms have not been fully understood yet (48). Moreover, cell type-specific opposite regulation of the same gene has been recently described for thyroid receptor, a nuclear receptor highly homologous to GR that has been shown to positively regulate the type 1 iodothyronine deiodinase gene in COS-7 and HEK-293 cells and suppress the same gene expression in the JEG-3 choriocarcinoma cell line. This negative regulation was demonstrated to require both thyroid receptor-DNA binding and the presence of a JEG-3-specific transcription factor, not present in COS-7 or HEK-293 cells (49).
GC are more than simple inhibitors of inflammation and immunity. Indeed, GC have been shown to enhance the expres-sion of scavenger receptors (e.g. MARCO and CD163) as well as complement components (e.g. C1q, C3, and C5), TLR family members, SAA family proteins, fibrinogen, the collectin MBL, and surfactant proteins SP-A and SP-D (50 -56). Interestingly, PTX3 binds C1q and activates the classic pathway of complement activation, and both are augmented by GC. SAA, MBL, fibrinogen, SP-A, and SP-D play key roles in the extracellular matrix, are components of the humoral arm of innate immunity produced by nonhematopoietic cells, and share with PTX3 the property to be augmented by GC.
PTX3 is a prototypic long pentraxin structurally related to the short pentraxins CRP and SAP (3,4,6). It was previously shown that induction of the acute phase proteins (e.g. SAA) is impaired in the liver of GR dim/dim mice (57). GC selectively induce regulate acute phase and extracellular matrix proteins, such as SAA family proteins, fibrinogen, and surfactant proteins SP-A and SP-D, but not CRP and SAP. In addition, GC oppositely regulate IL-6 (inhibition) and IL-10 (induction), key inducers of CRP/SAP and PTX3 (1,3).
PTX3 binds TSG-6 and acts as a focal point in the assembly of hyaluronic acid-rich extracellular matrices. PTX3 levels are augmented in humans after GC administration and are elevated in patients affected by Cushing syndrome. Interestingly, PTX3 is found to localize in the derma of these patients. Patients with Cushing syndrome have alterations in the extracellular matrix with loss of connective tissue, leading to fragility and atrophy of the skin, thinning of the stratum corneum, and loss of subcutaneous fat. All of this explains easy bruising of skin after minimal trauma, the frequent presence of extensive ecchymoses, and the formation of typical striae rubrae on the stretched, fragile skin. It remains to be ascertained whether unbalanced production of PTX3 by fibroblasts and endothelial cells induced by GC plays a role in these manifestations of Cushing syndrome.
The results presented here indicate that the regulation of PTX3 by GC is cell context-dependent, with inhibition in hematopoietic cells (macrophages and DC) and induction in nonhematopoietic cells (fibroblasts and endothelial cells). The divergent effects of GC on PTX3 production in diverse cellular contexts is likely to reflect the diverse roles of this multifunctional protein in innate immunity, inflammation, and in the assembly of the extracellular matrix.